Interesting beam design

[trainguy]

Guys,

I need to design a beam that has a very different ultimate bending strength when loaded downward versus when it is loaded upward. (Actually its downward versus sideways, but I'm simplifying)

Imagining a rectangular tube as a section, if we laser cut a notch in the portion of the section that undergoes compression, and insert a metal plate with an interference fit, and we make sure that the reduced section meets shear, combined stress and fatigue requirements, could this work?

I would be using the "spacer" in bearing, to replace the compressive stress in the section. Under tension from a load in the other direction, the beam would have a much reduced S, or at least that's the hope. Some tack welding would be used to ensure the adjacent parts do not buckle locally when undergoing the primary bending stresses.

What do you think?

tg

 

[hokie66]

The notch would cause a stress riser in the tension case. Sounds to me like a case for full scale testing of a prototype.

 

[msquared48]

Liberty Ship Hatches Revisited!

Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask

 

[dhengr]

Trainguy:
To answer your last question.... first.... not much, about what? Otherwise..... wow. If I had the vaguest idea what the hell that word picture said and represented, I would think it was a wonderful concept. Just don’t design and build my trains that way.

BA will be along shortly to suggest a sketch, so I won’t bother. Although, I do tend to agree with him more often than not. He has that phrase stored in his computer and just brings it up whenever needed.

I would question your “combined stresses and fatigue requirements” if you are using tack welds.

 

[BAretired]

Seems that the rectangular hollow section fits the unsimplified criteria without modification. With the greater dimension vertically, it is stronger downward than it is sideways. Isn't that what you wanted?

It would not work for your "simplification" because the downward and upward strength would be the same.

I think dhengr wants you to draw a sketch but is afraid to say so.

BA

 

[apsix]

If it was down vs. up it could work, but fatigue failure might be difficult to eliminate, depending on the nature of the loads.

I would be inclined to look at a bolted moment connection with endplates and small top bolts.

Since it's actually downward versus sideways why not look at a slender I beam?

 

[paddingtongreen]

Whatever method, do you want the failure loads to be predictable, I hope not.
Since dhengr and BAretired are too chicken to ask, could we please have a sketch.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[csd72]

paddingtongreen,

and ruin all the fun of trying to guess why someone would cut up a perfectly good beam?

 

[rscassar]

Why don't you use a rectangular hollow section with the strong axis orientated in the direction of greater load.

 

[Cadair]

You might look at wood. Glulams can be different both top, bottom, and side. Wood is different in all directions!

 

[a2mfk]

I'm entertained when an abstract post gets so many guesses with no further input from the OP...

 

[rb1957]

ideas coming from across the board ...

we need something to refine the suggestions ...

what are you planning on using ? Al (plane), St (large sclae building), wood (small building) ...

similarly, what application ?

how much difference ?

how optiminal do you want to be ? this sounds like a weight-limited thing. As suggested an open rectangle, maybe with doublers on the heavier loaded faces ... ???

 

[dhengr]

Thank you Paddington...., I hate to be the mean one every time, I was hoping to get BA to do my dirty work for me, and you had to finally do it. I might add that a better explanation of what he is really trying to accomplish would be helpful too, if he wants better answers.

I think we should all find out what Trainguy was smoking when he wrote the OP, because he is usually much more lucid and coherent than that. I can usually follow his train(?) of thought, but I don’t think the AAR would go along with that OP.

 

[trainguy]

OK, OK guys.

After I got blasted (by forum members) for apparently spending too much company time on Eng-tips (is there such a thing as too much time on ET...), I've limited my visits to 1/day.

I'll admit I sent the request out in a rather limited amount of time, and a sketch would have helped. dh, BA, it's OK to ask for a sketch, given that 80% of my posts are just that.

I'm surprised and impressed that I already got 12 responses in less than 24hrs.

This is for a coupler carrier on a train. Its a simply supported steel beam, usually U shaped in cross section, about 30 inches long, with springs resting inside the U.

The carrier has to meet traditional (80 yr old) requirements of resisting a large (100k) downward force at midspan, with no yield, but has to also fail at relatively low loads (hoping for <20k) horizontally, when this coupler retracts (new state of the art requirements) during a train accident!

I am toying with the idea of using aluminum, because of some failures that I have observed over the years, where out of plane, short duration loads punched through aluminum train bodies and ends.

The challenge is trying to design a beam that is strong vertically but that simply disappears in horizontal impacts.

As far as my original question - here's the sketch:

If I slice a loaf of bread with a very large knife but I only cut to 1/3 of its depth, and I keep the knife in the bread, does the bread still work as a beam?

tg

 

[hokie66]

trainguy,
See my first post.

 

[slickdeals]

I think the knife has no role in the strength of the bread unless the load is physically able to travel through the knife. It might in compression, but not in tension. As others have mentioned, you essentially have a notch.

 

[rb1957]

maybe a flanged U-channel. in normal bending, the base of the U is in tension, and the section is limited by the compression stress in the up-right flanges of the U; so it looks like you can design for the load with some careful sizing of the flanges.

transverse bending pits one side of the U against the other, pretty much all you've got to work with is the spacing between the sides.

how about sizing the section you need for the jobs it has to do, then adapting it locally for the springs; like maybe an I-beam with U pockets, or a U channel that splays out to create a larger pocket for the springs ??

now i expect you'll add that the springs are the full length of the piece in question ...

 

[trainguy]

msquared48,

Good reference - thanks to Wikipedia, I understand what you're talking about.

Yikes...

tg

 

[IRstuff]

Perhaps you might think of corrugation, which is quite strong in the axis parallel to the ruts, but has almost zero strength perpendicular.

Or, you could put a semi-locked hinge that would given way when something else breaks.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[dhengr]

Trainguy:
You didn’t get blasted for your “Who’s Paying for This Time” thread # 731-286515, you brought up a mighty relevant question, and a bunch of people had trouble rationalizing their screwin around on company time, so they just pointed back at the OP’er.

Do you suppose there is anyone here other than you and me who even know what a coupler carrier and coupler carrier springs, wear plates, etc. are? What about striker bars or striker castings and coupler horns, which prevent your 20k lateral failure load on the carrier unless they fail first? What kind of a coupler angle and striker mouth opening are you dealing with that you have a 30" long carrier? I haven’t subscribed to the AAR Circular Letters for years and my Car Construction Manuals are pretty old too, but where do these loads come from and how are they applied, don’t be so vague. Nest your edress in your answer to these questions, something, something at something dot whatever. Is the lateral load only in compression on the car, or in tension also? I don’t remember the AAR having a 100k down on the coupler carrier requirement. But, I do remember designing around that at some point in time.

Are you talking about transit equipment, locomotives or freight equipment? I’ll be damned...., I dug out a 30yr old NYCTA spec. and that’s probably where I remember the ‘100k down, on the carrier and its connections, for any horiz. position for the coupler, at less than yield.’ Not just centered on the carrier because that won’t give the worst end reactions. And, it’s a curved beam too isn’t it? Don’t mess around notching the carrier for fatigue and vibration reasons, etc. Doesn’t the anti-climber arrangement effectively prevent the 20k lateral load on the coupler carrier? You have to tell me how this load can be applied, what’s their thinking on this, if you could ever get a straight and meaningful answer from them. Maybe nest the carrier in the anti-climber and end sill structure so it can take the reactions caused by the 100k down loading. But then, bolt it into the stringers w.r.t. longitudinal loading so that the bolts shear off at 10 or 15k, and the whole carrier just gets pushed back into the car, but I can’t see your details from here. In fact, in some crude way you can sell this for more $$ by calling it EOC cushioning, since the bolts absorb energy as they shear off.